Add response validation rules

Add validation so study participants must enter an email formatted like name@school.edu for question 2, require ages to be whole numbers between 18 and 99 on question 5, and force the postcode field to be exactly five digits. If a value fails, show a clear error: 'Enter a valid school email', 'Age must be a whole number between 18 and 99', 'Postcode must be 5 digits.' Prevent submission until these are fixed.

Before I send this survey to 500 alumni, surface the most likely data-entry errors and make them hard to make. Highlight which questions accept free text, require numeric input, or must match an email pattern. Add real-time validation messages for common mistakes and flag any open-text questions that probably need a drop-down instead to improve consistency.

I released the intake form and now 30 responses have '198' or '202' in the birth-year field and emails like 'jsmith' without a domain. The people involved are me, the data analyst Priya, and the participants. I don't know whether to retroactively clean the data or stop new bad responses immediately. What's the most likely cause and the fastest sequence to stop more bad data while preserving current submissions for minimal cleanup?